ferm

2.7K posts

ferm banner
ferm

ferm

@thelongferment

riddled with bugs, needs regular burping

Katılım Mayıs 2021
1K Takip Edilen81 Takipçiler
ferm retweetledi
Siddhartha Saxena
Siddhartha Saxena@siddsax·
Anthropic onboarding day: Michael Scott introducing Karpathy like he just signed Wemby in free agency.
English
378
1.4K
16.9K
2.1M
ferm
ferm@thelongferment·
@mnolangray This is so true and I don’t see an obvious solution, technical or otherwise. You still need that fallible human in the loop
English
0
0
0
69
ferm retweetledi
M. Nolan Gray 🥑
M. Nolan Gray 🥑@mnolangray·
This is exactly right. As they exist today, AI agents are the prototypical bad employee who nonetheless pleases the boss: they will affirm your ideas without second thought, come up with elaborate plans for how others might do the work, insist that they don't need help and never admit to uncertainty, take shortcuts and falsify work, and produce output that looks plausible at the level that a busy manager can scrutinize, but that is basically flawed. That's why any decent manager knows that hiring the right candidate is 95% of the job of management. You need a candidate who actually cares about the work, or is at least sufficiently scared of the repeat game reputational consequences of poor work, that they will exhibit the occasional disagreeableness, vulnerability, or neuroticism needed to produce work that achieves anything. In our work writing (and passing) California housing law, even a single poorly written line of code, a single undefined term, a single incorrect cross-reference, can derail billion dollar interventions and waste untold hours of time. You need staff/consultants/advocates who will occasionally exhibit the superficially "bad" qualities mentioned above, traits that are annoying to bad managers, but are necessary to avoid disasters.
Aaron Levie@levie

CEOs are uniquely prone to AI psychosis because they’re sufficiently distant from the last mile of work that still has to happen to generate most value with AI. So when they play with AI, they see the happy path results, often not considering the next 10 or 20 things that have to happen to get sustainable results from agents. “Look I made this awesome product prototype”. Yes but you didn’t have to review the code before it went into production and fix a bunch of issues. “Look I generated a contract”. Yes but you didn’t verify all the terms before it goes out to the counterparty and didn’t have to wire up all the past contracts to work with. The best thing you can do as a CEO is to use AI a *ton* to figure out the real implications of agents in the enterprise, and come out the other side with an appreciation for both the upside and the real work that goes into them.

English
5
22
283
17.6K
ferm
ferm@thelongferment·
@tenobrus Here’s the free ChatGPT answer 😂😂 complete with fully hallucinated justification
ferm tweet mediaferm tweet mediaferm tweet media
English
0
0
4
138
ferm
ferm@thelongferment·
@Romy_Holland @alyssaleann I have a fond memory of eating a slice on the median out front when I was doing the Berkeley bakery tour years ago. One of the original “artisan” bread shops in the country!
English
0
0
0
41
Alyssa 🌻
Alyssa 🌻@alyssaleann·
I wonder if a restaurant that made roughly 4 healthy-ish and delicious things dining-hall style but cheap and not-modifiable (allergies and pickiness totally ignored) could make a go of it or if the cat is entirely out of the bag on what we expect of food these days Restaurants - the concept of showing up somewhere and ordering what you want - are fairly new. It would be interesting to see someone try to rewind the clock to "you get what you get, but hey, it's cheap!"
English
17
0
82
3.3K
Romy
Romy@Romy_Holland·
@alyssaleann there’s a pizza place in berkeley that sells one flavor of pizza per day. it’s wildly popular.
English
3
0
12
354
ferm retweetledi
Colin McCarthy
Colin McCarthy@US_Stormwatch·
Solar in 2025 grew 19x faster than experts at the International Energy Agency predicted in 2015. Solar is now the fastest growing electricity source in human history.
Colin McCarthy tweet media
English
77
347
1.3K
126.1K
ferm retweetledi
alz
alz@alz_zyd_·
AI will revolutionize pure math, essentially dominating human mathematicians within 5 years, but this won't move the needle on technological progress, because the vast majority of modern pure math is useless for any practical application
English
34
12
225
46.7K
ferm retweetledi
prerat
prerat@prerat·
if i reset your memory every session, i bet you would write the same cliches over and over too
English
8
5
199
2.8K
ferm
ferm@thelongferment·
@literalbanana Pure magic. The feeling of singing a beloved song and playing it on the guitar at the same time was absolutely intoxicating
English
0
0
2
60
Science Banana
Science Banana@literalbanana·
Do you remember when you first got interested in your hobby or special interest? What was it like?
English
42
4
101
15.1K
ferm
ferm@thelongferment·
@0interestrates Most of the time code is a means to an end while writing is the end itself. Also, writing is just wayyyyy closer to “what a human is” than code. We’re both more sensitive to the idea of an AI doing that and more attuned to noticing irregularities in it.
English
0
1
1
17
rahul
rahul@0interestrates·
why do people (including me) have an aversion to AI writing but not as much to AI code? if a piece of text smells AI i stop reading it but i use things coded entirely with AI every day
English
318
17
791
126K
ferm retweetledi
OpenAI
OpenAI@OpenAI·
Today, we share a breakthrough on the planar unit distance problem, a famous open question first posed by Paul Erdős in 1946. For nearly 80 years, mathematicians believed the best possible solutions looked roughly like square grids. An OpenAI model has now disproved that belief, discovering an entirely new family of constructions that performs better. This marks the first time AI has autonomously solved a prominent open problem central to a field of mathematics.
English
1K
3.8K
26.4K
13.1M